Project checklist

Question 1

1.

Answer 1

1. Frame the problem and look at the big picture

Question 2

2.

Answer 2

2. Get the data

Question 3

3.

Answer 3

3. Explore the data to gain insights

Question 4

4.

Answer 4

4. Prepare the data

Question 5

5.

Answer 5

5. Explore many different models and shortlist the best ones

Question 6

6.

Answer 6

6. Fine-tune your models and combine them into a great solution

Question 7

7.

Answer 7

7. Present your solution

Question 8

8.

Answer 8

8. Launch, monitor, and maintain your system

Question 9

1.1

Answer 9

1.1 Understand business problem and objective, explore hidden assumptions

Question 10

1.2

Answer 10

1.2 Current solution

Question 11

1.3

Answer 11

1.3 Measure of success

Question 12

2.1

Answer 12

2.1 Get access to data and create workspace

Question 13

2.2

Answer 13

2.2 Consider legal and privacy obligations

Question 14

2.3

Answer 14

2.3 Set test data aside

Question 15

3.1

Answer 15

3.1 Check feature characteristics

Question 16

3.2

Answer 16

3.2 Visualize the data

Question 17

3.3

Answer 17

3.3 Study feature correlations

Question 18

4.1

Answer 18

4.1 Data cleaning

Question 19

4.2

Answer 19

4.2 Feature selection and engineering

Question 20

4.3

Answer 20

4.3 Feature scaling

Question 21

5.1

Answer 21

5.1 Train and compare many quick and dirty models

Question 22

5.2

Answer 22

5.2 Quick feature selection and engineering, iterate

Question 23

5.3

Answer 23

5.3 Shortlist 3-5 most promising models, preferring different errors

Question 24

6.1

Answer 24

6.1 Fine tune hyper-parameters, optimize against business objective

Question 25

6.2

Answer 25

6.2 Try ensemble method

Question 26

6.3

Answer 26

6.3 Measure performance against test set

Question 27

7.1

Answer 27

7.1 Document

Question 28

7.2

Answer 28

7.2 Explain achievement of business objective

Question 29

7.3

Answer 29

7.3 Present other interesting points

Question 30

8.1

Answer 30

8.1 Make production ready

Question 31

8.2

Answer 31

8.2 Write monitoring code

Question 32

8.3

Answer 32

8.3 Set up retraining of model